Forensic Science International 295 (2019) 226.e1–226.

e9

Contents lists available at ScienceDirect

Forensic Science International

journal homepage: www.elsevier.com/locate/forsciint

Forensic Anthropology Population Data

Testing and comparing the accuracy of two dental age estimation

methods on saudi children: Measurements of open apices in teeth and
the London Atlas of Tooth Development
Dara M. Alsudairi* , Sakher J. AlQahtani
Department of Pediatric Dentistry and Orthodontics, College of Dentistry, King Saud University, Riyadh, 11545, Saudi Arabia

A R T I C L E I N F O A B S T R A C T

Article history: Evaluation of dental development is frequently performed to assess maturity and estimate age. The aim
Received 29 January 2018 of this study was to test and compare the accuracy of two dental age estimation methods in a sample of
Received in revised form 26 September 2018 Saudi children by using Cameriere’s formula (measurements of mandibular teeth open apices) and The
Accepted 13 November 2018
London Atlas of Tooth Development. Four hundred archived digital panoramic radiographs of healthy
Available online 22 November 2018
children (200 males and 200 females) between the ages of 6 to 15.99 years attending the Pediatric Dental
clinics of King Saud University, College of Dentistry, Riyadh, Saudi Arabia were used. The mean difference
Keywords:
between the Estimated Age (EA) and Real Age (RA) was ( 0.89 years) for Cameriere’s formula with a
Age estimation
Forensic anthropology population data
standard deviation of (1.14 years), and it was ( 0.59 years) with a standard deviation of (1.45 years) for
Forensic dentistry The London Atlas. Both methods consistently underestimated the age. No statistically significant
Saudi Arabia difference between the sexes was found in both methods In conclusion, the London Atlas was found to be
Cameriere’s method slightly more accurate and easier to use than was Cameriere’s formula.
London Atlas © 2018 Published by Elsevier B.V.

1. Introduction considered challenging to apply in living individuals, as they

necessitate the presence of extracted teeth, and are time-
Age estimation, in the absence of proper documents, is becoming consuming.
increasingly important and is a basic human right [1]. It answers New methods of non-invasive dental age estimation have
questions of law, when age is disputed, in areas such as human emerged to overcome some of the previous limitations. These
identification and asylum seeking, given the escalating dilemma of include Cameriere’s formula [9] and The London Atlas of Tooth
immigrants, criminal liability, and child abuse [2]. The importance of Development [10].
age estimation in identifying human bodies was evident after the In 2006, Cameriere et al. measured the chronological age of
occurrence of several natural and mass disasters worldwide. children in relation to the open apices of roots [9]. They studied
Moreover, it helps monitor the growth and development of children different populations in Europe (Italy, Germany, Croatia, Slovenia,
and guides decisions for starting certain types of treatment. Kosovo, UK, and Spain) and developed a standard formula in 2007
Dental development is generally preferred in age estimation that has also been tested on different populations [11].
due to its low variability as the calcification of dental structures is The seven permanent mandibular teeth on the left side are
regulated by genes instead of environmental factors [3–5]. In assessed using Cameriere’s formula. Teeth with completed root
addition, they survive inhumation well and develop over an development (closed apices) are measured (N0). Developing teeth
extended period of the human life with measurable stages. with open apices are measured by calculating the distance between
Multiple odontogenic methods were introduced to primarily the inner sides of the open apex (A). In teeth with two roots, the sum
assess the mineralization stages of Tooth Development. distances between inner sides of both open apices are calculated. To
Dental age estimation methods can be classified into two broad eliminate any magnification, the measurements are divided by tooth
categories: invasive and non-invasive. Invasive techniques include length (L) (normalized measurements). Dental maturity, therefore, is
aspartic acid racemization, radiocarbon analysis, and incremental evaluated as the sum of the normalized open apices (s) and the
lines in the root cementum [6–8]. In general, these methods are number of teeth with complete root development (N0), and is
replaced in the following regression formula for age estimation:
Age = 8.971 + 0.375 g + 1.631  5 + 0.674 N0 1.034 s 0.176 s N0,
* Corresponding author.
E-mail address: d.alsudairi@gmail.com (D.M. Alsudairi). where g stands for sex (1 = males and 0 = females).

https://doi.org/10.1016/j.forsciint.2018.11.011
0379-0738/© 2018 Published by Elsevier B.V.
 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9 226.e2

In 2010, AlQahtani et al. developed a comprehensive atlas of representation, and proper representation of the dentition and
Tooth Development to estimate age between 28 weeks intrauter- osseous structures. Only patients of Saudi nationality with no
ine and 23 years by examining mixed Bangladeshi and white systemic or developmental problems were included in this study.
British individuals [10] by using Moorrees et al.’s and Bengston’s Finally, the presence of all permanent teeth, excluding the third
Tooth Developmental stages [12,13]. molars, in both dentitions of the left side was mandatory.
A limited number of studies in the area of forensic dentistry were
conducted and published in the past few years in Saudi Arabia. 2.3. Exclusion criteria
Demirjian’s method of dental age estimation was the most
commonly used by Saudi researchers [14–17]. Few studies examined Poor-quality radiographs that exhibited dark areas (burnout),
the London Atlas of Tooth Development and demonstrated multiple light areas, an overlap of structures, tooth size discrepancy, and
weaknesses [18,19]. Therefore, further verifications and investiga- artifacts were excluded from the study. Subjects with any systemic
tions are needed in this field and should involve more representative diseases, developmental problems, and abnormal dental develop-
samples and valid statistical analyses that would enable us to ment were also omitted. Other conditions such as the presence of
compare the Saudi population with other populations. gross pathology related to the jaw or teeth, gross caries, periapical
This study had several main aims. The first aim was to pathosis, and large restorations or crowns were also excluded.
determine the accuracy of Cameriere’s formula and the London Subjects with early tooth extraction and previous orthodontic
Atlas of Tooth Development in estimating the chronologic age of treatment were also excluded.
Saudi pediatric patients who seek care at the College of Dentistry,
King Saud University, Riyadh, Saudi Arabia. The second aim was to 2.4. Measurements
investigate whether a difference in the accuracy of age estimation
existed between Cameriere’s formula and the London Atlas of When The London Atlas was used, digital radiographs were
Tooth Development. The third aim was to examine whether a assessed by direct comparison with the diagrams and by using the
difference existed between the estimated age of male and female dental age estimator feature in the London Atlas online software
patients when using Cameriere’s formula and the London Atlas of [23]. A digital radiograph software was used for tooth measure-
Tooth Development. ments using Cameriere’s method [24]. The chronological age of
We proposed three hypotheses. The first hypothesis relates to each subject was calculated by subtracting the date of the
Saudi children, where the null hypothesis suggests that there will radiograph from the date of birth registered in the file and was
be no difference between the chronological age and the age converted into a decimal system by using Eveleth and Tanner’s
estimated using Cameriere’s formula and the London Atlas of Tooth method [25]. The chronological ages were blinded from the
Development. The second hypothesis is associated with the age observer and entered in a different spreadsheet until the
estimation method, where the null hypothesis indicates that there assessment of all 400 panoramic radiographs was completed.
will be no difference in the age estimation accuracy of Cameriere’s The main observer was trained by an expert on a set of separate
formula and the London Atlas of Tooth Development. The last radiographs to ensure the correct execution of Cameriere’s
hypothesis is about sex, where the null hypothesis implies that measurements before initiating the study.
there will be no difference in the accuracy of Cameriere’s formula
and the London Atlas of Tooth Development between male and 2.5. Methods of age estimation
female subjects.
2.5.1. Cameriere’s method
2. Materials and methods All the left mandibular permanent teeth (excluding the third
molars) were assessed. In single-rooted teeth, the distance
2.1. Sample selection and size between the inner side of the pulp canal at the apex was measured
(Ai, i = 1, . . . ,5) [9]. In teeth with two roots, the sum of the
This was a retrospective cross-sectional study on children distances between the inner sides of the pulp of the two apices was
between the ages six and 15.99 years. Each chronological year was assessed (Ai, i = 6, 7) [9]. To rule out any possible variation in
assigned as an individual group. At significance level α = 0.05, magnification and angulation of the radiographs, the measure-
estimated standard deviation (SD) = 0.9, maximum difference = 0.8, ments were normalized by dividing the measurement by the
and power = 81%, the sample size in each group was calculated to length of the tooth from the highest cusp tip to the lowest root tip
be 40, which was then divided to include 20 male and 20 female (Li, i = 1, . . . ,7) [9] (Fig. 1). Dental age was estimated by entering
subjects per group. Therefore, 400 panoramic radiographs the normalized measurements (xi = Ai/Li, i = 1, . . . ,7), the sum of
(200 males and 200 females) were utilized. the normalized open apices (s = x1 + x2 + x3 + x4 + x5 + x6 + x7), and
Ethical approval was granted by the College of Dentistry’s the number of teeth with complete root development (N0) in the
Research Center at King Saud University (registration number PR following formula:
0050). All radiographs utilized in this project were archived at the
Dental Age = 8.971 + 0.375g + 1.631x5 + 0.674N0
Pediatric Dental Department and were acquired as part of the
1.034s 0.176s.N0,
routine dental treatment of patients; no radiographs were chosen
specifically for this project. where g stands for sex (males = 1 and females = 0), x5 = A5/L5,
The radiographs were collected by entering the inclusion s = sum of normalized open apices, and N0 = number of teeth with
variables, and this yielded approximately 1000 radiographs complete root development [9].
acquired between the years 2015 and 2017. The final sample size
was calculated to be 400, which was comparable to those of 2.5.2. The London Atlas of human Tooth Development and eruption
previously published papers in the literature [15,20–22]. Direct comparison of the radiographs with the diagrams of The
London Atlas of human Tooth Development and the dental age
2.2. Inclusion criteria calculator feature in the London Atlas software were used to
estimate the dental age. The permanent dentition of the left side of
Only good-quality radiographs were included with adequate both the jaws was used. Tooth Developmental stages were also
exposure, maximum contrast, optimum density, exact anatomical tabulated.
226.e3 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9

correlation yielded a Cohen’s kappa of 0.89, which indicated

excellent reliability as well.

3.2. Description of the sample

Four hundred radiographs were assessed (200 males and 200

females). All subjects were divided into ten groups based on their
chronological age, with each group having equal numbers of males
and females (20 each). The groups included different age ranges as
follows: group 1 (6.00–6.99), group 2 (7.00–7.99), group 3
(8.00–8.99), group 4 (9.00–9.99), group 5 (10.00–10.99), group 6
(11.00–11.99), group 7 (12.00–12.99), group 8 (13.00–13.99), group
9 (14.00–14.99), and group 10 (15.00–15.99).

3.3. Cameriere’s method

Fig. 1. Cameriere’s measurements of open apices (De Luca et al.). (For interpretation
of the references to colour in this figure legend, the reader is referred to the web
3.3.1. The accuracy of the entire sample
version of this article.)
Blue arrow: A in a single-rooted tooth. The mean chronological age of the entire sample was
Red arrows: A in a multi-rooted tooth. 10.89 years. The mean estimated age of the whole sample
Green arrow: L (tooth length). determined using Cameriere’s method was 10.00 years. The
difference of 0.89 years (SD  1.14 years) was significant
(P  0.05) and indicated a consistent underestimation. The bias
2.6. Intraobserver and interobserver agreement
was statistically significant in all the age groups (Table 1).

The primary observer was trained and calibrated by an expert in

3.3.2. The accuracy for male and female subjects
the field. All measurements were later performed by the main
The results for male and female subjects were compared using
observer. The interobserver and intraobserver reproducibility were
an independent-samples t-test (Table 2). No statistically significant
assessed using a random sample of 10% of the radiographs and by
difference (P = 0.269) was observed between male and female
reassessing them after an interval of at least two weeks by using a
subjects, and the mean difference between the chronological and
kappa test [26].
estimated ages was less in female subjects than in male subjects.

2.7. Statistical analysis

3.3.3. The accuracy of age intervals of the entire sample
The paired t-test results for the entire sample and each year
Data were analyzed using IBM SPSS Statistics for Windows/
interval are illustrated in Fig. 2. Age was significantly
Macintosh, Version 22.0 (IBM Corp., Armonk, NY, USA). For
overestimated only in age group 6, while it was significantly
Cameriere’s method, all the morphological variables, Xi, s N0,
underestimated in most age groups, including groups 8, 10, 11, 12,
and sex were entered. The values were kept in the linear regression
13, 14, and 15.
equation formulated by Cameriere et al. for their study on
European populations [11].
3.3.4. The accuracy of age intervals for male and female subjects
Dental age (Estimated age) was subtracted from the chrono-
The paired t-test results for male and female subjects for each
logical age (Real age): a positive result indicated overestimation
year interval are shown in Fig. 3. While non-significant,
and a negative one indicated underestimation. Bias was the mean
overestimation was noticed for both sexes in age group 6 and
difference between the estimated and real ages determined using a
paired t-test. That was used for each of the two methods with a Table 2
significance level of P < 0.05. The accuracy of Cameriere’s method for male and female subjects expressed by bias.
The absolute mean difference between the estimated and real
Independent-samples test
ages was calculated to express the accuracy independent of bias.
Measure of accuracy Sex N Mean SD SE Significance
mean
3. Results
Bias F 200 0.8100 1.06116 0.07504 0.269
M 200 0.9366 1.21943 0.08623
3.1. Reliability tests
The absolute F 200 1.0325 0.84497 0.05975 0.047
difference M 200 1.0325 0.84497 0.05975
For Cameriere’s method, the intraobserver comparison yielded
*F = Female.
a Cohen’s kappa of 0.91, which indicated excellent reliability. For *M = Male.
the London Atlas of Tooth Development, the intraobserver *SD = Standard deviation.
comparison yielded a Cohen’s Kappa of 0.84, and the interobserver *SE = Standard error.

Table 1
The accuracy of Cameriere’s method for the entire sample expressed by bias.

Measure of accuracy N Mean SD SE mean Significance 95% CI lower 95% CI upper

Bias 400 0.8933 1.14337 0.05717 0.000 0.9857 0.7609
The absolute difference 400 1.1217 0.90021 0.04501 0.000 1.0332 1.2102

*SD = Standard deviation.

*SE = Standard error.
*CI = Confidence interval.
Bold value means the result is significant as mentioned above the table when P< or equal to 0.05.
 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9 226.e4

Fig. 2. The difference between estimated and real age for Cameriere’s method.

Fig. 3. The bias of age groups using Cameriere’s method for male and female subjects.
226.e5 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9

for male subjects in age group 7. Underestimation was uniform Table 4

The accuracy of the London Atlas for male and female subjects expressed by bias.
among sexes in all the other age groups.
Independent-samples test
3.4. The London Atlas of human Tooth Development Measure of accuracy Sex N Mean SD SE Significance
mean
3.4.1. The accuracy of the entire sample Bias F 200 0.4306 1.44462 0.10215 0.029
The mean chronological age of the entire sample was M 200 0.7475 1.45206 0.10268
10.89 years. The mean estimated age of the whole sample The absolute F 200 1.0898 1.03899 0.07347 0.169
determined using the London Atlas was 10.31 years. The difference difference M 200 1.2349 1.06650 0.07541

of 0.59 years (SD  1.45 years) was significant (P  0.05) and *F = Female
indicated a consistent underestimation. The bias was statistically *M = Male
SD = Standard deviation
significant in all the age groups (Table 3).
*SE = Standard error

3.4.2. The accuracy for male and female subjects

The results for male and female subjects were compared using The proper evaluation of dental development for estimating the
an independent-samples t-test (Table 4). No statistically significant age of living individuals involves the use of radiographs, which
difference (P = 0.029) was observed between male and female despite the effects of radiation exposure are still very reliable and
subjects, and the mean difference between the chronological and can be used for treatment. This consistency has been validated and
estimated ages was less in female subjects than in male subjects. has helped in identifying disaster victims [30].
The use of The London Atlas of Tooth Development is based on
3.4.3. The accuracy of age intervals of the entire sample identifying the correct stage of Tooth Development, which relies on
The paired t-test results for the entire sample and each year the knowledge, precision, and training of the examiner. Therefore,
interval are illustrated in Fig. 4. Age was significantly both intraobserver and interobserver reliability tests are done.
overestimated only in age group 6 and non-significantly in groups However, Cameriere’s method is a very objective technique based
7, 8, and 9. Age was significantly underestimated in the other on measurements; therefore, only an intraobserver analysis is
groups, including groups 10, 11, 12, 13, 14, and 15. performed.

3.4.4. The accuracy of age intervals for male and female subjects 4.1. The accuracy of the entire sample
The paired t-test results for male and female subjects for each
year interval are shown in Fig. 5. While non-significant, Several studies in the literature tested Cameriere’s formula on
overestimation was noticed for both sexes in age groups 6, 7, different populations following the same protocol, which made the
and 8 and for female subjects in age group 9. Underestimation was results comparable. This method consistently underestimated the
uniform among sexes in all the other age groups. age, albeit to varying degrees (Indian population: 0.18 years;
Iranian population: 0.18 years; Turkish population: 0.35 years;
4. Discussion and American population: 1.25) [31–34]. Although the Saudi
population also showed underestimation, the difference was more
The identity of an individual is determined by many pronounced ( 0.87 years).
fundamental aspects, including name, sex, nationality, and, more Studies involving the Mexican and Bosnian Herzegovinian
importantly, age. Surprisingly, 30–50% of births are still unregis- populations did not reveal the general bias [20,35]. Nevertheless,
tered [27]. The failure to register birth to authorities leads to the they revealed different results between the sexes.
child not having an identity the society, and this is in violation of Underestimation of age was a common finding in several
the child’s human rights declared in article seven of the UN studies using the London Atlas. A study conducted by AlQahtani
Convention on the Rights of the Child [1]. et al. that included a wide range of ethnicities underestimated the
Social interviews and physical developmental graphs have age by 0.10 years [36]. Another study on an American population
been favored for use amongst authorities. However, they harbor exhibited an underestimation of 0.17 years [34]. These results
many weaknesses ranging from not being thoroughly tested to concur with the results of the current research, but the
the fact that physical development is highly influenced by underestimation is more pronounced in the Saudi sample
environmental factors [28]. For example, the hand and wrist age ( 0.59 years).
assessment requires a radiograph that cannot be used for Overestimation was noticed in two different studies conducted
treatment; its purpose is solely for screening. Therefore, it may in Portugal. This overestimation was manifested by 0.1 years and
be considered a disadvantage because it exposes patients to 1 month, respectively [37,38].
unnecessary radiation [29]. A study conducted by AlQahtani et al. in 2017 simultaneously
In contrast, dental development is the least influenced by compared three different populations. The samples came from
environmental factors and has been extensively studied and tested. Saudi Arabia, Spain, and Italy. The results indicated no significant

Table 3
The accuracy of the London Atlas for the entire sample expressed by bias.

One-sample statistics

Measure of accuracy N Mean SD SE Mean Significance 95% CI lower 95% CI upper

Bias 400 0.5890 1.45521 0.07276 0.000 0.7321 0.4460
The absolute difference 400 1.1624 1.05401 0.05270 0.000 1.0588 1.2660

*SD = Standard deviation.

*SE = Standard error.
*CI = Confidence interval.
Bold value means the result is significant as mentioned above the table when P< or equal to 0.05.
D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9 226.e6

Fig. 4. The difference between estimated and real age for the London Atlas.

Fig. 5. The bias of age groups using the London Atlas for male and female subjects.
226.e7 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9

difference among populations and exhibited no bias (Saudi In the Mexican population, the difference between both the
population: 0.247 years; Spanish population: 0.099 years; and sexes and age intervals was <1 year. An exception was observed for
Italian population: 0.451 years) [39]. female subjects in the 14- to 15-year age group wherein an
In Saudi Arabia, a study published by Alshihri et al. stated that underestimation of 1.06 years was apparent [20].
the majority of the sample (65.5%) did not exceed the 12-month In the American population, the results were divided
range. Meanwhile, the percentage of overestimation by more than according to ethnicities. For Americans of Indian descent, the
12 months was 19%, and underestimation was 15.5% [18]. In their underestimation exceeded one year in the following two age
methodology, instead of comparing the real age to the estimated groups (13 and 15+) [34]. European populations showed both
age, they compared their means. Additionally, their sample size underestimation and overestimation by more than one year. The
was lower (n = 252) than that in this study (n = 400). Moreover, overestimation was displayed in the age groups 6–7, 7–8, and
their sample did not have a uniform distribution between male and 9–10 years [34]. The underestimation was exclusively observed in
female subjects. Finally, the age range of their sample was more the age group 17–18 years [34]. Hispanic populations exhibited
extensive as they included subjects aged from 2 to 20 years. All of overestimation in the 8- to 9-year age group, while the 16+ year
these aspects make it difficult to conduct a complete comparison age group showed underestimation [34].
between the two studies. In Iran, the estimation was the most accurate in males aged
Another study by Alshihri et al. that compared several age 6–12 years and in females aged 6–11 years [32].
estimation methods found that the London Atlas underestimated In Brazil, although Cameriere’s method was proven accurate,
the age by 1.6 (15.0) months [19]. overestimation was noticed in females aged between 5 and
All the cited differences in the results of both the methods may 10 years and underestimation was noticed in those aged between
be attributed to uncontrolled elements such as biological factors. 11 and 15 years [40].
These differences may also be attributed to components of the The results of this study indicate a constant general
research conducted. For example, the massive sample size in the underestimation that was more pronounced in the last age
study by AlQahtani et al. can make it difficult to establish a category of 15–15.99 years, with a mean difference of 2.168 years.
comparison with studies that included a smaller number of The only overestimation was observed in the age group
subjects in their study and with less uniform distribution. 6–6.99 years, with a mean difference of 0.398 years.
Regarding sex, female subjects exhibited overestimation only in
4.2. The accuracy for male and female subjects the age group 6–6.99 years, whereas male subjects exhibited
overestimation in two age groups: 6–6.99 years and 7–7.99 years.
The underestimation of age was uniform between the sexes in A significant difference was not observed between most age
several studies that used Cameriere’s formula (Indian population: categories when using the London Atlas in the American
males = 0.11 years and females = 0.23 years; Iranian population: population [35]. The age categories that stood out were 6, 10,
males = 0.27 years and females = 0.11 years; Turkish population: 15, 16, and 17 years, with mean values of 1.25, 0.33, 0.48, 1.16,
males = 0.74 years and females = 0.24 years; and American and 1.05 years, respectively [35].
population: males = 1.23 years and females = 1.19 years) [31–34]. In the study by Alshihri et al. on western Saudi children, a
This coincides with the findings of this study, which found no significant difference between the EA and RA was found in the age
significant difference between male and female subjects, even groups 7–9 years and 13–15 years [18].
though the difference was more pronounced (males = 0.94 years A year later, they conducted a comparison study on integrating
and females = 0.81 years). standard age estimation methods in western Saudi children. The
Other studies reported a difference in age estimation between London Atlas was found to be the most accurate in the age group
the sexes. Overestimation in females was common in two 7–9 years, wherein 78.3% of subjects were within the 12-month
populations (Mexican population: 0.10 years and Bosnian range [19]. For the age group 16–18 years, the London Atlas was the
Herzegovinian population: 0.09 years) [20,35]. In males, the results least accurate as 50% of subjects were within the normal range
differed with an accurate estimation (0.00) in the Mexican sample [19]. The dental age was overestimated in 56.3% of subjects aged
and an underestimation of 0.02 years in the Bosnian between 4 and 6 years, while it was underestimated in 78.3% of
Herzegovinian sample [20,35]. those aged between 7 and 9 years [19].
One study confirmed there was no significant difference AlQahtani et al. found no significant difference between most
between bias in male and female subjects when using the London age categories of the London Atlas when it was compared with the
Atlas (American population, P > 0.005) [34]. The results of this Schour and Massler estimation chart and Ubelaker estimation
research agree with those of the previous one that found chart. However, exceptions were noted for the age groups 3, 4, 16,
no significant difference between the sexes (bias for male 19, 20, 21, 22, and 23 years, with mean values of 0.12, 0.24, 0.43,
subjects = 0.75 years and for female subjects = 0.43 years). 0.54, 0.76, 0.82, 1.59, and 1.83 years, respectively [36].
Overestimation of age in male subjects was noticed in two other The results of this study indicate a significant difference
studies (Cesario et al. = 2.4047 and Pavlovic et al. = 2.1296) [37,38]. between the sample in general in each age category (underesti-
However, age was underestimated in female subjects in only two of mation). However, a significant difference was not observed
those studies [36,37], while showing no significance in the study between male and female subjects.
conducted by Pavlovic et al. [38]. The hypotheses proposed in this study tackle three aspects: the
difference between the chronological and estimated ages, the
4.3. The accuracy of age intervals of the entire sample and male and accuracy of both age estimation methods in general, and accuracy of
female subjects age estimation between the sexes. Our proposed null hypothesis was
that there would not be any difference between them. These results
Not all studies presented the results regarding age intervals as proved that the only difference existed between the chronological
most of them focused only on the general associations. Below is a and estimated ages in both techniques. No differences were observed
comparison between the available sources in the literature. in accuracy either between the methods in general or between male
In the Turkish population, a general finding was that the and female subjects under both approaches.
differences between the EA and RA decreased when the RA Although the results of this study, in conjunction with those of
increased for each age group when using Cameriere’s method [33]. several universal studies, indicate an underestimation of age, the
 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9 226.e8

accuracy does not seem to be affected. Cameriere’s measurements [7] K.L. Spalding, B.A. Buchholz, L.E. Bergman, H. Druid, J. Frise ’, Forensics: age
provide a point estimate that suggests a real difference between written in teeth by nuclear tests, Nature 437 (2005) 333–334, doi:http://dx.
doi.org/10.1038/437333a.
the estimated and real ages. For example, 7.5 implies 7 years and [8] P. Aggarwal, S. Saxena, P. Bansal, Incremental lines in root cementum of human
6 months. In the London Atlas, age is expressed in a range rather teeth: an approach to their role in age estimation using polarizing microscopy,
than as a definite number. For instance, 7.5 represents the mean of Indian J. Dent. Res. 19 (2008) 326–330, doi:http://dx.doi.org/10.4103/0970-
9290.44536.
an age range from 7.00 to 7.99 years. This indicates that the [9] R. Cameriere, L. Ferrante, M. Cingolani, Age estimation in children by
difference between EA and RA could be exaggerated. measurement of open apices in teeth, Int. J. Legal Med. 120 (2006) 49–52, doi:
On the basis of previous facts and the outcomes of other http://dx.doi.org/10.1007/s00414-005-0047-9.
[10] S.J. AlQahtani, M.P. Hector, H.M. Liversidge, Brief communication: the London
extensive studies, a conclusion was drawn regarding the accuracy Atlas of human tooth development and eruption, Am. J. Phys. Anthropol. 142
of dental age estimation methods. The age is considered exact (2010) 481–490, doi:http://dx.doi.org/10.1002/ajpa.21258.
depending on how close it is from any given range. In other words, [11] R. Cameriere, D. De Angelis, L. Ferrante, F. Scarpino, M. Cingolani, Age
estimation in children by measurement of open apices in teeth: a European
if a confidence interval of 2 years around the estimated age exists,
formula, Int. J. Legal Med. 121 (2007) 449–453, doi:http://dx.doi.org/10.1007/
it is considered accurate [41]. s00414-007-0179-1.
[12] C.F. Moorrees, E.A. Fanning, E.E. Hunt Jr., Age variation of formation stages for
5. Conclusion ten permanent teeth, J. Dent. Res. 42 (1963) 490–502, doi:http://dx.doi.org/
10.1177/00220345630420062701.
[13] R.G. Bengston, A study of the time of eruption and root development of the
In general, the London Atlas of Tooth Development proved permanent teeth between six and thirteen years, Dent. Res. Grad Study 35
slightly more accurate than did Cameriere’s method, even though (1935) 3–9.
[14] M. Ajmal, K.I. Assiri, K.Y. Al-Ameer, A.M. Assiri, M. Luqman, Age estimation
the difference was not significant. In contrast, a considerable using third molar teeth: study on southern Saudi population, J. Forensic Dent.
difference was observed between the chronological and estimated Sci. 4 (2012) 63–65, doi:http://dx.doi.org/10.4103/0975-1475.109886.
ages in both techniques, which indicated a consistent underesti- [15] Z.D. Baghdadi, Dental maturity in Saudi children using the Demirjian method:
a comparative study and new prediction models, ISRN Dent. 2013 (2013)
mation. Furthermore, no differences were observed in the
390314, doi:http://dx.doi.org/10.1155/2013/390314.
measured variables between male and female subjects in both [16] Z.D. Baghdadi, S.C. Pani, Accuracy of population-specific Demirjian curves in
methods. the estimation of dental age of Saudi children, Int. Paediatr. Dent. 22 (2012)
The London Atlas of Tooth Development was easier to use for 125–131, doi:http://dx.doi.org/10.1111/j.1365-263X.2011.01179.x.
[17] A. Alshihri, E. Kurger, M. Tennant, Dental age assessment of 4-16-year-old
several reasons. First and foremost, it is less time-consuming than Western Saudi children and adolescents using Demirjian’s method for forensic
are the linear measurements conducted under Cameriere’s dentistry, Egypt J. Forensic Sci. 6 (2016) 152–156, doi:http://dx.doi.org/
method. Another point is that teeth in both the jaws can be used 10.1016/j.ejfs.2015.03.003.
[18] A. Alshihri, E. Kurger, M. Tennant, Dental age assessment of Western Saudi
rather than the ones in the left lower jaw alone. In addition, both children and adolescents, Saudi Dent. J. 27 (2015) 131–136, doi:http://dx.doi.
dentitions can be used instead of just the permanent ones used in org/10.1016/j.sdentj.2015.01.002.
Cameriere’s technique. Finally, the availability of an extra dental [19] A. Alshihri, E. Kurger, M. Tennant, Integrating standard methods of age
estimation in western Saudi children and adolescent, Eur. J. Forensic Sci. 3
developmental stage in the London Atlas that details the scores of (2016) 157–162, doi:http://dx.doi.org/10.5455/ejfs.201936.
apex closure results in better measures of accuracy. [20] S. De Luca, S. De Giorgio, A.C. Butti, R. Biagi, M. Cingolani, R. Cameriere, Age
estimation in children by measurement of open apices in tooth roots: study of
Declarations of interest a Mexican sample, Forensic Sci. Int. 221 (2012) 155, doi:http://dx.doi.org/
10.1016/j.forsciint.2012.04.026.
[21] Z. Kırzıoglu, D. Ceyhan, Accuracy of different dental age estimation methods on
None. This research was part of a Master’s Degree requirement Turkish children, Forensic Sci. Int. 216 (2012) 61–67, doi:http://dx.doi.org/
and did not receive any specific grant from funding agencies in the 10.1016/j.forsciint.2011.08.018.
[22] P. Sharma, V. Wadhwan, R. Prakash, S. Goel, P. Aggarwal, Age estimation in
public, commercial, or not-for-profit sectors. children by measurement of open apices in teeth: a study in North Indian
population, Aust. J. Forensic Sci. 48 (2016) 592–600, doi:http://dx.doi.org/
CRediT authorship contribution statement 10.1080/00450618.2015.1112426.
[23] Atlas of Tooth Development and Eruption, https://atlas.dentistry.qmul.ac.uk/
2016. (Accessed 1 March 2016).
Dara M. Alsudairi: Data curation. Sakher J. AlQahtani: Project [24] Planmeca Romexis 3.6.0. R, https://www.planmeca.com/Software/news-
administration. videos/releases/2016. (Accessed 1 March 2016).
[25] P.B. Eveleth, J.M. Tanner, Worldwide Variation in Human Growth, second ed.,
Cambridge University Press, New York, 1990, pp. 397.
[26] M.L. McHugh, Interrater reliability: the kappa statistic, Biochem. Med. 22
Acknowledgements
(2012) 276–282, doi:http://dx.doi.org/10.11613/BM.2012.031.
[27] UNICEF, Factsheets: Birth Registration, (2012) http://www.unicef.org/news-
The authors would like to thank the college of Dentistry at King line/2003/03fsbirthregistration.htm. (Accessed 1 May 2012).
Saud University, Riyadh, Saudi Arabia, for providing the facilities, [28] D.T. Wigle, Child Health and the Environment, Oxford University Press, New
York, 2003.
equipment and personnel that helped preforming this research. [29] S. Dhiman, S. Maheshwari, S.K. Verma, Assessment of maturity in orthodon-
tics: a review, J. Adv. Clin. Res. Insights 2 (2015) 100–103, doi:http://dx.doi.org/
10.15713/ins.jcri.54.
References [30] A. Alonso, P. Martin, C. Albarrán, P. Garcia, L. Fernandez de Simon, M. Jesús
Iturralde, A. Fernández-Rodriguez, I. Atienza, J. Capilla, J. García-Hirschfeld, P.
[1] Children’s Rights Alliance, Summary of the UN Convention on the Rights of the Martinez, G. Vallejo, O. García, E. García, P. Real, D. Alvarez, A. León, M. Sancho,
Child, (2013) https://www.childrensrights.ie/sites/default/files/information_- Challenges of DNA profiling in mass disaster investigations, Croat. Med. J. 46
sheets/files/SummaryUNCRC.pdf. (Accessed 12 February 2018). (2005) 540–548.
[2] T. Solheim, A. Vonen, Dental age estimation, quality assurance and age [31] A. Shrestha, R.P. Yadav, S. Shrestha, I.K. Maharjan, S. Camelio, Measurement of
estimation of asylum seekers in Norway, Forensic Sci. Int. 159 (Suppl. 1) (2006) open apices in teeth for estimation of age in children, Health Renaissance 12
S56–S60, doi:http://dx.doi.org/10.1016/j.forsciint.2006.02.016. (2014) 33–37, doi:http://dx.doi.org/10.3126/hren.v12i1.11983.
[3] C. Nolla, The development of permanent teeth, J. Dent. Child. 27 (1960) [32] S. Javadinejad, H. Sekhavati, R. Ghafari, A comparison of the accuracy of four
254–266. age estimation methods based on panoramic radiography of developing teeth,
[4] C.F. Moorrees, E.A. Fanning, E.E. Hunt Jr., Age variation of formation stages for J. Dent. Res. Dent. Clin. Dent. Prospects 9 (2015) 72–78, doi:http://dx.doi.org/
ten permanent teeth, J. Dent. Res. 42 (1963) 1490–1502, doi:http://dx.doi.org/ 10.15171/joddd.2015.015.
10.1177/00220345630420062701. [33] A. Gulsahi, R.E. Tirali, S.B. Cehreli, S. De Luca, L. Ferrante, R. Cameriere, The
[5] I. Gleiser, E.E. Hunt, The permanent mandibular first molar; its reliability of Cameriere’s method in Turkish children: a preliminary report,
calcification, eruption, and decay, Am. J. Phys. Anthropol. 13 (1995) Forensic Sci. Int. 249 (2015) 319, doi:http://dx.doi.org/10.1016/j.for-
253–283, doi:http://dx.doi.org/10.1002/ajpa.1330130206. sciint.2015.01.031.
[6] P.M. Helfman, J.L. Bada, Aspartic acid racemization in tooth enamel from living [34] S.A. Santana, J.D. Bethard, T.L. Moore, Accuracy of dental age in non-adults: a
humans, Proc. Natl. Acad. Sci. U. S. A. 72 (1975) 2891–2894. comparison of two methods for age estimation using radiographs of
226.e9 D.M. Alsudairi, S.J. AlQahtani / Forensic Science International 295 (2019) 226.e1–226.e9

developing teeth, J. Forensic Sci. 62 (2017) 1320–1325, doi:http://dx.doi.org/ [38] S. Pavlovic, C. Palmela Pereira, R.F. Vargas de Sousa Santos, Age estimation in
10.1111/1556-4029.13434. Portuguese population: the application of the London atlas of tooth
[35] I. Gali
c, M. Vodanovic, R. Cameriere, E. Nakaš, E. Gali
c, E. Selimovi
c, H. Brki
c, development and eruption, Forensic Sci. Int. 272 (2017) 97–103, doi:http://
Accuracy of Cameriere, Haavikko, and Willems radiographic methods on dx.doi.org/10.1016/j.forsciint.2017.01.011.
age estimation on Bosnian-Herzegovian children age groups 6-13, Int. J. [39] S.J. AlQahtani, J. Adserias, E. Nuzzolese, G. Di Vella, The accuracy of the London
Legal Med. 125 (2011) 315–321, doi:http://dx.doi.org/10.1007/s00414-010- Atlas of human tooth development and eruption in dental age estimations of
0515-8. Saudi, Spanish, and Italian children, Conference Paper: Annual Scientific
[36] S.J. AlQahtani, M.P. Hector, H.M. Liversidge, Accuracy of dental age estimation Meeting of the American Academy of Forensic Science (2017) 69.
charts: Schour and Massler, ubelaker and the London Atlas, Am. J. Phys. [40] M.M. Fernandes, R.L. Tinoco, D.P. de Braganca, S.H. de Lima, L. Francesquini
Anthropol. 154 (2014) 70–78, doi:http://dx.doi.org/10.1002/ajpa.22473. Junior, E. Daruge Junior, Age estimation by measurements of developing teeth:
[37] C. Cesário, R. Santos, D. Pestana, C. Palmela Pereira, Medico-legal age accuracy of Cameriere’s method on a Brazilian sample, J. Forensic Sci. 56 (2011)
estimation in a sub-adult Portuguese population: validation of Atlas Schour 1616–1619, doi:http://dx.doi.org/10.1111/j.1556-4029.2011.01860.x.
and Massler and London, J. Civil Legal Sci. 5 (2016) 196, doi:http://dx.doi.org/ [41] L. Kullman, Accuracy of two dental and one skeletal age estimation method in
10.4172/2169-0170.1000196. Swedish adolescents, Forensic Sci. Int. 75 (1995) 225–236.